Abstract--We prove that the marginals densities of a primal normal factor graph and the corresponding marginal densities of its dual normal factor graph are related via local mappings. The mapping relies on no assumptions on the size, on the topology, or on the parameters of the graphical model. The mapping provides us with a simple procedure to transform simultaneously the estimated marginals from one domain to the other, which is particularly useful when such computations can be carried out more efficiently in one of the domains. In the case of the Ising model, valid configurations in the dual normal factor graph of the model coincide with the terms that appear in the high-temperature series expansion of the partition function. The subgraphs-world process (as a rapidly mixing Markov chain) can therefore be employed to draw samples according to the global probability mass function of the dual normal factor graph of ferromagnetic Ising models.

In this paper, we explore the effectiveness of dynamic analysis techniques for identifying malware, using Hidden Markov Models (HMMs) and Profile Hidden Markov Models (PHMMs), both trained on sequences of API calls. We contrast our results to static analysis using HMMs trained on sequences of opcodes, and show that dynamic analysis achieves significantly stronger results in many cases. Furthermore, in contrasting our two dynamic analysis techniques, we find that using PHMMs consistently outperforms our analysis based on HMMs.

Compressive sensing (CS) has been studied and applied in structural health monitoring for wireless data acquisition and transmission, structural modal identification, and spare damage identification. The key issue in CS is finding the optimal solution for sparse optimization. In the past years, many algorithms have been proposed in the field of applied mathematics. In this paper, we propose a machine-learning-based approach to solve the CS data-reconstruction problem. By treating a computation process as a data flow, the process of CS-based data reconstruction is formalized into a standard supervised-learning task. The prior knowledge, i.e., the basis matrix and the CS-sampled signals, are used as the input and the target of the network; the basis coefficient matrix is embedded as the parameters of a certain layer; the objective function of conventional compressive sensing is set as the loss function of the network. Regularized by l1-norm, these basis coefficients are optimized to reduce the error between the original CS-sampled signals and the masked reconstructed signals with a common optimization algorithm. Also, the proposed network can handle complex bases, such as a Fourier basis. Benefiting from the nature of a multi-neuron layer, multiple signal channels can be reconstructed simultaneously. Meanwhile, the disassembled use of a large-scale basis makes the method memory-efficient. A numerical example of multiple sinusoidal waves and an example of field-test wireless data from a suspension bridge are carried out to illustrate the data-reconstruction ability of the proposed approach. The results show that high reconstruction accuracy can be obtained by the machine learning-based approach. Also, the parameters of the network have clear meanings; the inference of the mapping between input and output is fully transparent, making the CS data reconstruction neural network interpretable.

Linear mixture models have proven very useful in a plethora of applications, e.g., topic modeling, clustering, and source separation. As a critical aspect of the linear mixture models, identifiability of the model parameters is well-studied, under frameworks such as independent component analysis and constrained matrix factorization. Nevertheless, when the linear mixtures are distorted by an unknown nonlinear functions -- which is well-motivated and more realistic in many cases -- the identifiability issues are much less studied. This work proposes an identification criterion for a nonlinear mixture model that is well grounded in many real-world applications, and offers identifiability guarantees. A practical implementation based on a judiciously designed neural network is proposed to realize the criterion, and an effective learning algorithm is proposed. Numerical results on synthetic and real-data corroborate effectiveness of the proposed method.

Acoustic scene classification is the task of identifying the scene from which the audio signal is recorded. Convolutional neural network (CNN) models are widely adopted with proven successes in acoustic scene classification. However, there is little insight on how an audio scene is perceived in CNN, as what have been demonstrated in image recognition research. In the present study, the Class Activation Mapping (CAM) is utilized to analyze how the log-magnitude Mel-scale filter-bank (log-Mel) features of different acoustic scenes are learned in a CNN classifier. It is noted that distinct high-energy time-frequency components of audio signals generally do not correspond to strong activation on CAM, while the background sound texture are well learned in CNN. In order to make the sound texture more salient, we propose to apply the Difference of Gaussian (DoG) and Sobel operator to process the log-Mel features and enhance edge information of the time-frequency image. Experimental results on the DCASE 2017 ASC challenge show that using edge enhanced log-Mel images as input feature of CNN significantly improves the performance of audio scene classification.

Though neither inherently good nor evil on its own, AI is powerful. It's how we choose to use AI that says more about us than it does the technology. What does an AI future look like to you? Today, China is home to an estimated 200 million surveillance cameras, about four times as many as in the United States. Scanning the faces of citizens in an effort to catch wanted criminals, police are able to identify individuals from drug smugglers to jaywalkers and collecting more information along the way.

This week's plunge in Apple's shares was another sign of the impact of the relations between the US and China. It does look like President Trump's tariffs are taking a toll and that the tensions maybe lasting. So this why Kai-Fu Lee's book, AI Superpowers: China, Silicon Valley, and the New World Order, is so timely. He provides a detailed look at how China is poised to win one of the most important markets. Keep in mind that – according to a research report by PWC – AI is forecasted to add $15.7 trillion to global GDP by 2030. The main reason is that this technology is general purpose, having applications that span industries like healthcare, transportation, financial services, energy and so on.

Alibaba DAMO Academy, the global research program launched by Alibaba in 2017, has published its predictions for the Top 10 technology trends in 2019. From speech AI, super-large graph neural networks, heterogenous computing architecture, to autonomous driving, blockchain and data protection technologies, machine intelligence has been generating great impacts on our lives, and an accelerated pace of technology revolution is expected in the year ahead. Real-time urban simulation becomes possible More resources will be allocated to technologies powering an intelligent "city brain" and its applications, while a city simulation model reflecting the real-time impulses and movements of a physical city can be built to facilitate the optimisation of city governance. More cities in China are expected to have a "city brain" in 2019. Speech AI in certain areas to pass Turing Test As speech intelligence technology advances, realtime text-to-speech on mobile devices would be almost identical to human speech, even passing the Turing test in certain conversations, such as ones using a robotic voice to alert about delivery status.

Martin Hiesboeck is an international business consultant with a focus on companies' operations in Asia.

The rise of automation has been reviewed by the World Bank chief economist based on data collated from a number of industries. Most advanced economies seen a decline in industrial jobs since the year 2000 and a rise of robots, in other parts of the world, notably East Asia, there has been a net gain of manufacturing jobs and little sign of robots replacing these types of roles. Other predictions have been less optimistic. For example, in 2017 Oxford University researchers Dr. Michael Osborne and Dr. Carl Frey interpreted data which suggested that over fifty percent of jobs in a developed economy are vulnerable in terms of humans being replaced by machines. Similarly, the World Economic Forum forecasts that machines and automated software will be handling fully half of all workplace tasks by 2025.